Sign up for our newsletter and get the latest HPC news and analysis.

MacWorld looks at OpenCL

MacWorld takes a look at OpenCL in this article at its web site. You may recall from earlier coverage that OpenCL is to GPGPU programming as MPI was to early message passing machines

…Nvidia calls its technology Compute Unified Device Architecture (CUDA) and AMD, through its ATI subsidiary, calls its technology Close To Metal (CTM). Both accomplish the same thing: They let programmers write code that enables them to send data to the GPU to be processed, rather than having to rely on the CPU.

Not content to yield the performance graphics market to competitor AMD, Intel is preparing its “Larabee” microprocessor for release in 2009.

…With no end in sight to the constant escalation of competing graphics technologies from AMD and Nvidia—and with Intel now acting as a wildcard—all bets are off at this point. So developing platform-neutral GPGPU technology only makes sense. And that seems to be where Apple’s focus is with OpenCL.

Why MacWorld? Apple is pushing OpenCL through the Khronos Group (the fine folks who manage OpenG, etc.) as a major feature of the next rev of its OS.

AMD has also announced support for OpenCL, and last week announced that it would be extending its Stream SDK to support it

AMD logoAs previously announced AMD is also supporting efforts to develop OpenCL as an open standard and plans to evolve the Stream SDK to be OpenCL compliant. Through equal support for both DirectX 11 and OpenCL, and by continuing to give developers the option of creating and using their own programming languages and high level tools, AMD is executing on a strategy designed to give programmers maximum choice and flexibility.

Resource Links: